Iwamatsu-san, > > It recognizes NOR flash in 5.10.y. How about checking DTS? > > Sodia board uses n25q512a and has the following DT: > > ``` > > &qspi { > > status = "okay"; > > > > flash0: n25q512a@0 { > > #address-cells = <1>; > > #size-cells = <1>; > > compatible = "n25q512a"; > > reg = <0>; > > spi-max-frequency = <100000000>; > > > > m25p,fast-read; > > cdns,page-size = <256>; > > cdns,block-size = <16>; > > cdns,read-delay = <4>; > > cdns,tshsl-ns = <50>; > > cdns,tsd2d-ns = <50>; > > cdns,tchsh-ns = <4>; > > cdns,tslch-ns = <4>; > > };t > > }; > > ``` > I am going to look into the difference. I confirmed that "n25q512a" in the node name and the compatible string are the only difference. Interestingly, if I used compatible = "n25q512a" instead of compatible = "s25fl128s1" or "s25fl128s", which is the current setting for linux-4.19.177-cip, the kernel prints the following message and is able to detect spi-nor flash as expected. [ 2.701294] spi-nor spi0.0: found s25fl128s1, expected n25q512a [ 2.711093] spi-nor spi0.0: s25fl128s1 (16384 Kbytes) Probably the detect mechanism has been changed by the commit. Apparently setting compatible = "n25q512a" for my board does not make sense. I'd like to find out why compatible = "s25fl128s1" does not work. Regards, Takuo Koguchi > -----Original Message----- > From: 小口琢夫 / KOGUCHI，TAKUO > Sent: Monday, March 15, 2021 2:31 PM > To: cip-dev@lists.cip-project.org > Subject: RE: [cip-dev] Is cadence-quadspi working on cyclone V? > > Iwamatsu-san, > > Thank you for the information. > > It recognizes NOR flash in 5.10.y. How about checking DTS? > > Sodia board uses n25q512a and has the following DT: > > ``` > > &qspi { > > status = "okay"; > > > > flash0: n25q512a@0 { > > #address-cells = <1>; > > #size-cells = <1>; > > compatible = "n25q512a"; > > reg = <0>; > > spi-max-frequency = <100000000>; > > > > m25p,fast-read; > > cdns,page-size = <256>; > > cdns,block-size = <16>; > > cdns,read-delay = <4>; > > cdns,tshsl-ns = <50>; > > cdns,tsd2d-ns = <50>; > > cdns,tchsh-ns = <4>; > > cdns,tslch-ns = <4>; > > }; > > }; > > ``` > I am going to look into the difference. > > Takuo Koguchi > > > -----Original Message----- > > From: cip-dev@lists.cip-project.org On > > Behalf Of Nobuhiro Iwamatsu > > Sent: Monday, March 15, 2021 2:25 PM > > To: cip-dev@lists.cip-project.org > > Subject: [!]Re: [cip-dev] Is cadence-quadspi working on cyclone V? > > > > Hi Koguchi-san, > > > > > -----Original Message----- > > > From: cip-dev@lists.cip-project.org > > > [mailto:cip-dev@lists.cip-project.org] On Behalf Of Takuo Koguchi > > > Sent: Monday, March 15, 2021 1:33 PM > > > To: cip-dev@lists.cip-project.org > > > Subject: Re: [cip-dev] Is cadence-quadspi working on cyclone V? > > > > > > Hi Pavel, > > > > > > > >> Git bisect has told me the first bad commit is > > > > >> 31fb632b5d43 spi: Move cadence-quadspi driver to drivers/spi/ > > > > > > > > > > That's strange. > > > > > > > > > >> Does this change requires device tree modification? > > > > >> Any suggestions or comments would be appreciated. > > > > > > > > > > ...because that particular change does not change any C code. > > > > > But it touches Kconfig bits; can you check that > > > > > CONFIG_SPI_CADENCE_QUADSPI is enabled in the "bad" configuration? > > > > > > > > > > Do you have any local changes in > > > > > drivers/mtd/spi-nor/controllers/cadence-quadspi.c or > > > > > drivers/spi/spi-cadence-quadspi.c ? > > > > I have no local changes to those files. > > > > Maybe I lost CONFIG_SPI_CADENCE_QUADSPI accidentally while bisecting. > > > > I will try bisect and post the result again. > > > > > > > > > > I redo the bisect and get the following result; # first bad commit: > > > [a314f6367787ee1d767df9a2120f17e4511144d0] mtd: spi-nor: Convert > > > cadence-quadspi to use spi-mem framework > > > > > > It contains significant amount of changes; > > > > > > cadence-quadspi.c | 476 > > > +++++++++++++++++++++--------------------------------- > > > 1 file changed, 191 insertions(+), 285 deletions(-) > > > > > > With this change, cpspi_probe return 0 without any error messages, > > > but it does not detect NOR flash on a custom cyclone V board. > > > > > > Any information will be appreciated. > > > > I checked on a Sodia board with the same SoC. > > > > https://secure-web.cisco.com/1ChOMrKnSBR5QZeejvrNs3Pf4x8__M_QVUHErnB9s > > lV-TAWP3NUwe0LEzsirvM7rr-Oqv > > RKrgxgI3iQo5ITdUehFyDppd2GljkEzx7izv16pERniX0DeDmzwE5h9pUYqtIxO3WAFTDu > > H_EG-RNhNFqK4LQakZ90sESF6 > > DMvs_jovN55Xj5u9kNOC_ew76pLtAMAOdN3BZSV_fR_n6tcWlhvqJ6OlUEfrpz9xvy-gu1 > > VHrgcgj1BICaBid9bVUza1-T7rBc > > I2EjECtd0qCMcsvLYZPxvzfZ9-Hu6rFlaCQ0ruhxZVfHuVWEB37eQmm0HLHYJG2_R_ousV > > eVDk8bJa2iA/https%3A%2F%2 > > Fgit.kernel.org%2Fpub%2Fscm%2Flinux%2Fkernel%2Fgit%2Fcip%2Flinux-cip.g > > it%2Ftree%2Farch%2Farm%2Fboot%2Fd > > ts%2Fsocfpga_cyclone5_sodia.dts%3Fh%3Dlinux-5.10.y-cip > > > > It recognizes NOR flash in 5.10.y. How about checking DTS? > > Sodia board uses n25q512a and has the following DT: > > ``` > > &qspi { > > status = "okay"; > > > > flash0: n25q512a@0 { > > #address-cells = <1>; > > #size-cells = <1>; > > compatible = "n25q512a"; > > reg = <0>; > > spi-max-frequency = <100000000>; > > > > m25p,fast-read; > > cdns,page-size = <256>; > > cdns,block-size = <16>; > > cdns,read-delay = <4>; > > cdns,tshsl-ns = <50>; > > cdns,tsd2d-ns = <50>; > > cdns,tchsh-ns = <4>; > > cdns,tslch-ns = <4>; > > }; > > }; > > ``` > > > > > > Thanks, > > > > > > Takuo Koguchi > > > > > > > Best regards, > > Nobuhiro > > > > > > > > -----Original Message----- > > > > From: 小口琢夫 / KOGUCHI，TAKUO > > > > Sent: Sunday, March 14, 2021 4:37 PM > > > > To: cip-dev@lists.cip-project.org > > > > Subject: Re: [!]Re: [cip-dev] Is cadence-quadspi working on cyclone V? > > > > > > > > Hi Pavel, > > > > Thank you for the response. > > > > > > > > >> Git bisect has told me the first bad commit is > > > > >> 31fb632b5d43 spi: Move cadence-quadspi driver to drivers/spi > > > > > > > > > > That's strange. > > > > > > > > > >> Does this change requires device tree modification? > > > > >> Any suggestions or comments would be appreciated. > > > > > > > > > > ...because that particular change does not change any C code. > > > > > But it touches Kconfig bits; can you check that > > > > > CONFIG_SPI_CADENCE_QUADSPI is enabled in the "bad" configuration? > > > > > > > > > > Do you have any local changes in > > > > > drivers/mtd/spi-nor/controllers/cadence-quadspi.c or > > > > > drivers/spi/spi-cadence-quadspi.c ? > > > > I have no local changes to those files. > > > > Maybe I lost CONFIG_SPI_CADENCE_QUADSPI accidentally while bisecting. > > > > I will try bisect and post the result again. > > > > > > > > Regards, > > > > > > > > Takuo Koguchi > > > > > > > > > 2021/03/13 8:07、Pavel Machek のメール: > > > > > > > > > > Hi! > > > > > > > > > >> Git bisect has told me the first bad commit is > > > > >> 31fb632b5d43 spi: Move cadence-quadspi driver to drivers/spi/ > > > > > > > > > > That's strange. > > > > > > > > > >> Does this change requires device tree modification? > > > > >> Any suggestions or comments would be appreciated. > > > > > > > > > > ...because that particular change does not change any C code. > > > > > But it touches Kconfig bits; can you check that > > > > > CONFIG_SPI_CADENCE_QUADSPI is enabled in the "bad" configuration? > > > > > > > > > > Do you have any local changes in > > > > > drivers/mtd/spi-nor/controllers/cadence-quadspi.c or > > > > > drivers/spi/spi-cadence-quadspi.c ? > > > > > > > > > > Best regards, > > > > > Pavel > > > > > -- > > > > > DENX Software Engineering GmbH, Managing Director: Wolfgang Denk > > > > > HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, > > > > > Germany > > > > > > > > > > > > > > >